Full Code of SanshruthR/CCTV_YOLO for AI

Repository: SanshruthR/CCTV_YOLO
Branch: main
Commit: 4745dbecaec6
Files: 4
Total size: 5.0 KB

Directory structure:
gitextract_f2uacro9/

├── README.md
├── app.py
├── requirements.txt
└── yolov5n6.pt

================================================
FILE CONTENTS
================================================

================================================
FILE: README.md
================================================
# Fast Real-time Object Detection with High-Res Output

![Python](https://img.shields.io/badge/Python-3.8%2B-3776AB?style=for-the-badge&logo=python&logoColor=white)
![YOLOv5](https://img.shields.io/badge/YOLOv5n6-6.1-orange?style=for-the-badge&logo=YOLO&logoColor=white)
![CUDA](https://img.shields.io/badge/CUDA-Enabled-76B900?style=for-the-badge&logo=nvidia&logoColor=white)
![OpenCV](https://img.shields.io/badge/OpenCV-3.4.2-5C3EE8?style=for-the-badge&logo=opencv&logoColor=white)
![Gradio](https://img.shields.io/badge/Gradio-3.1.4-blueviolet?style=for-the-badge&logo=gradio&logoColor=white)
[![Deployed on Hugging Face](https://img.shields.io/badge/Deployed%20on-Hugging%20Face-yellow?style=for-the-badge&logo=huggingface&logoColor=white)](https://huggingface.co/spaces/Sanshruth/CCTV_YOLO?embed=true)
![image](https://github.com/user-attachments/assets/2cbae46f-e9de-48db-96f8-d439536ad703)


## Overview

Live stream of Las Vegas sidewalk traffic cam, processed with YOLOv5n6 on low-resolution frames, with results drawn on high-resolution frames.

This project demonstrates **real-time object detection** using the YOLOv5n6 model with **low-resolution inference** for high-speed processing, while drawing the results on **high-resolution frames**. The object detection pipeline is deployed as a Gradio app and streams live data from an external camera feed.

### Features

- **YOLOv5n6 Model**: Pre-trained object detection model optimized for speed and accuracy.
- **Low-resolution Processing**: Efficiently processes frames in low resolution (320x180) while mapping results to high-res images.
- **Gradio Interface**: Interactive Gradio interface with real-time video stream processing.
- **CUDA Support**: Optimized for CUDA-enabled GPUs, ensuring fast inference times.

### Model Details

- **Model**: YOLOv5n6 (`yolov5n6.pt`)
- **Confidence Threshold**: 0.25
- **IOU Threshold**: 0.45
- **Max Detections**: 100 objects per frame

### How It Works

The pipeline processes a live video stream, detecting objects in each frame using YOLOv5n6. Bounding boxes are drawn on the high-resolution frames based on detections from the low-resolution inference.

### Usage

1. Clone the repository and install the dependencies:
    ```bash
    git clone https://github.com/SanshruthR/CCTV_YOLO.git
    cd cctv-yolo
    pip install -r requirements.txt
    ```

2. Run the script:
    ```bash
    python app.py
    ```

3. Access the Gradio interface and view the live stream processed with YOLOv5n6.

### Deployment

This project is deployed on **Hugging Face Spaces**. You can interact with the app via the following link:

[Live Demo on Hugging Face](https://huggingface.co/spaces/Sanshruth/CCTV_YOLO?embed=true)

### License

This project is licensed under the MIT License.


================================================
FILE: app.py
================================================
from ultralytics import YOLO
import cv2
import numpy as np
import gradio as gr
import torch

# Load YOLOv5n6 model
model = YOLO('yolov5n6.pt')

# Set the confidence threshold and IOU
model.conf = 0.25  # confidence threshold
model.iou = 0.45  # IOU threshold
model.agnostic = False
model.multi_label = False
model.max_det = 100  # max number of detections

# Low-resolution for inference 
#use frame_count += 3
#LOW_RES = (500, 600) 

#Faster inference but less accurate 
#LOW_RES = (320, 180) 

#Fastest inference but low accuracy, change frame_count to += 2
LOW_RES = (300, 300)

def detect_and_draw(frame):
    # Create low-res copy
    low_res_frame = cv2.resize(frame, LOW_RES)
    
    # Perform detection
    results = model(low_res_frame, verbose=False)

    # Scale bounding boxes
    scale_x = frame.shape[1] / LOW_RES[0]
    scale_y = frame.shape[0] / LOW_RES[1]

    # Draw bounding boxes on high-res frame
    for detection in results[0].boxes.data:
        x1, y1, x2, y2, conf, cls = detection
        x1, y1, x2, y2 = int(x1*scale_x), int(y1*scale_y), int(x2*scale_x), int(y2*scale_y)
        label = f"{results[0].names[int(cls)]} {conf:.2f}"
        cv2.rectangle(frame, (x1, y1), (x2, y2), (0, 255, 0), 2)
        cv2.putText(frame, label, (x1, y1 - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.9, (0, 255, 0), 2)

    return frame

# Define your stream URL
stream_url = "https://edge01.london.nginx.hdontap.com/hosb5/ng_showcase-coke_bottle-street_fixed.stream/chunklist_w464099566.m3u8"

def process_stream():
    cap = cv2.VideoCapture(stream_url)
    frame_count = 0
    while cap.isOpened():
        ret, frame = cap.read()
        if not ret:
            break

        frame_count += 2  #change to 3 for higher res
        if frame_count % 30 == 0:
            result = detect_and_draw(frame)
            result_rgb = cv2.cvtColor(result, cv2.COLOR_BGR2RGB)
            yield result_rgb

    cap.release()

iface = gr.Interface(
    fn=process_stream,
    inputs=None,
    outputs="image",
    live=True,
    title="Fast Real-time Object Detection with High-Res Output",
    description="Live stream processed with YOLOv5n6 on low-res frames, results drawn on high-res frames."
)

if __name__ == "__main__":
    if torch.cuda.is_available():
        model.to('cuda')
    #iface.queue()
    iface.launch()


================================================
FILE: requirements.txt
================================================
ultralytics
opencv-python
gradio
torch